Autoimmune diseases, such as multiple sclerosis (MS) and type 1 diabetes (T1D), are caused by autoreactive T cells that have escaped negative selection, and which are potentially cross-reactive with mutant or foreign proteins. Similarly, anti-tumor immunity may be directed against antigens which are non-mutated self proteins, or which incorporate tumor-associated mutations into the recognized epitope. Until recently, structural studies of TCR/peptide/MHC complexes had been limited to anti- foreign TCRs. In 2005, however, we reported the structure of an autoimmune TCR bound to a self- peptide from myelin basic protein (MBP) and HLA-DR2a. Subsequently, we determined the trimolecular structure of a human melanoma-specific TCR recognizing a naturally-occurring somatic mutation in the glycolytic enzyme triosephosphate isomerase (mutTPI) in the context of HLA-DR1. Remarkably, these structures revealed unconventional binding topologies that appear suboptimal for TCR binding, and that may reflect the distinct selection pressures exerted on autoreactive as compared to anti-foreign TCRs. These results open the way to a systematic investigation of the structural and biophysical principles governing self-recognition in autoimmunity and cancer, and to using this knowledge to engineer agents to specifically modulate (suppress or enhance) T cell responses to self-antigens. Our objectives are: 1. Basis for TCR recognition of self-antigens. We will extend our previous work to determine whether DR4-restricted TCRs from MS and T1D patients also bind self-peptide/MHC with altered topologies and suboptimal interactions. 2. Basis for TCR recognition of tumor antigens. To compare recognition of non- mutated versus mutant tumor antigens, and interpret this in the context of self-antigen recognition in autoimmune diseases, we will examine TCR recognition of the non-mutated HLA-DR4-restricted melanoma antigens tyrosinase (Ty) and gp100. 3. High-affinity TCRs as potential immunotherapeutics for MS. Directed evolution (yeast display) will be used to engineer high-affinity MBP-specific TCRs as novel agents for targeting immunosuppressive cytokines to sites of autoantigen presentation. 4. Altered peptide ligands for enhancing anti-melanoma T cell responses. Based on binding and structural information on TCR recognition of Ty and gp100, these shared melanoma epitopes will be modified for heightened immunogenicity. 5. Ligand-induced TCR dimerization as a possible T cell signaling mechanism. As demonstrated in solution, the melanoma-specific TCR G4 dimerizes upon binding mutTPI/DR1. To test the hypothesis that the (G4/mutTPI/DR1)2 complex represents a basic T cell signaling unit, we will determine its structure, and assess the structure through mutagenesis and correlative functional analyses of T cell activation. Taken together, these studies will provide a comprehensive view of the biophysical basis for anti-self immunity in autoimmune diseases and cancer. Public Health Relevance Statement: While remarkable progress has been made in understanding the molecular basis for TCR recognition of microbial antigens, much less is known about the principles governing TCR recognition of self or altered self in autoimmune diseases and cancer. Our objective is to elucidate these principles through structural, biophysical and correlative functional analyses of autoimmune and tumor-specific TCRs, and to use this knowledge to engineer agents to specifically suppress or enhance T cell responses to self.